Statistics and Data Sciences

Dear Colleagues,

As we advance into the era of large-scale, data-centric research, fields like Large Vision Models (LVMs), Multimodal Image Processing, 3D Computer Vision, and Robotics Vision have become central to modern computer vision and data science. These domains require sophisticated statistical and data science methodologies to analyze and interpret extensive, high-dimensional datasets, enabling breakthroughs in fields like autonomous robotics, augmented reality, and multimodal AI. This Special Issue aims to explore the role of mathematical and statistical foundations in developing and optimizing models for these cutting-edge applications.

We invite high-quality research contributions on topics such as:

• Large Vision Models (LVMs): Statistical techniques for developing and refining LVMs, including model parameterization, uncertainty estimation, and interpretable modeling. This section also welcomes contributions on scaling LVMs for cross-domain applications and improving model generalization.
• Multimodal Image Processing: Techniques for integrating and processing data from multiple modalities (e.g., visual, textual, and auditory) to enhance image interpretation and scene understanding. This includes methods for multimodal feature fusion, as well as applications in complex tasks like cross-modal retrieval and multimodal event detection.
• 3D Computer Vision: Statistical and data science approaches to 3D image processing, 3D scene reconstruction, object tracking, and depth estimation. We particularly encourage submissions that address challenges in large-scale 3D data analysis and applications to fields like augmented reality and digital twins.
• Robotics Vision: Vision systems and statistical methodologies for robotic perception, navigation, and interaction in dynamic environments. Topics of interest include sensor fusion, visual SLAM (Simultaneous Localization and Mapping), and reinforcement learning techniques for visual tasks in robotics.
• Multimodal Deep Neural Network Model Development: Statistical and computational approaches to designing and training multimodal deep neural networks (DNNs), with a focus on model robustness, efficiency, and accuracy across different data modalities.
• High-Dimensional Data Analysis and Dimensionality Reduction: Methods for feature extraction, selection, and dimensionality reduction in high-dimensional vision and multimodal datasets. This includes computationally efficient techniques for handling large-scale complex data structures.
• Machine Learning and Statistical Learning Theory: Innovations in statistical learning for vision tasks, including Bayesian and non-parametric methods, and advancements in unsupervised, self-supervised, and transfer learning techniques.
• Big Data Computing and Algorithms: Scalable algorithms and computational frameworks designed for large-scale image, video, and multimodal data. This includes distributed processing, parallel computing approaches, and model architectures optimized for big data.
• Temporal Analysis and Forecasting Models: Statistical models for analyzing temporal patterns in video data and sequential image frames, including event prediction, activity recognition, and dynamic scene analysis.
• Applications in Biomedicine, Economics, and Social Sciences: Practical applications of vision and multimodal processing methods in medical imaging, biometric analysis, social interaction analysis, and economic data modeling, where statistical and data science techniques provide critical insights.

This Special Issue seeks to bridge theoretical research and practical applications, encouraging submissions that showcase innovative statistical methods and data science techniques that advance the fields of computer vision and multimodal AI. Through this collection, we aim to offer new perspectives on the integration of multimodal data, advanced vision models, and statistical approaches, driving the future of intelligent and autonomous systems.

Dr. Yanbing Bai
Dr. Jinze Yu
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Mathematics is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• large vision models
• multimodal image processing
• 3D computer vision